TDOA Source Localization in the Presence of Synchronization Clock Bias and Sensor Position Errors

Abstract

TDOA positioning requires the sensors to be synchronized when sampling the signals to locate an emitting source. It also demands the sensor positions to be accurate. Achieving synchronization may not be feasible when the sensors are widely separated or are many, and precise sensor positions are often not available in practice. This paper investigates TDOA localization using a set of imperfect sensors, where timing synchronization offsets and random position errors exist among different sensor groups. The degradation in localization accuracy caused by synchronization offsets and position errors is examined through the Cramér–Rao lower bound (CRLB) analysis under Gaussian noise. A novel method in obtaining sequentially algebraic solutions of the source location, the sensor positions, and the synchronization offsets is developed. Sequential estimation simplifies the localization task and offers complexity reduction. The performance of the proposed estimates is shown to reach the CRLB accuracy analytically using small noise analysis, although they are estimated sequentially and not jointly. Simulations support and corroborate the theoretical developments.